TK-BOM-Project/backend/app/services/instrument_classifier.py

"""
INSTRUMENT 분류 시스템 (간단 버전)
완제품 구매용 기본 분류만
"""

import re
from typing import Dict, List, Optional
from .material_classifier import classify_material

# ========== 기본 계기 타입 ==========
INSTRUMENT_TYPES = {
    "PRESSURE_GAUGE": {
        "dat_file_patterns": ["PG_", "PRESS_G"],
        "description_keywords": ["PRESSURE GAUGE", "압력계", "PG"],
        "characteristics": "압력 측정용 게이지"
    },
    "TEMPERATURE_GAUGE": {
        "dat_file_patterns": ["TG_", "TEMP_G"],
        "description_keywords": ["TEMPERATURE GAUGE", "온도계", "TG", "THERMOMETER"],
        "characteristics": "온도 측정용 게이지"
    },
    "FLOW_METER": {
        "dat_file_patterns": ["FM_", "FLOW_"],
        "description_keywords": ["FLOW METER", "유량계", "FM"],
        "characteristics": "유량 측정용"
    },
    "LEVEL_GAUGE": {
        "dat_file_patterns": ["LG_", "LEVEL_"],
        "description_keywords": ["LEVEL GAUGE", "액위계", "LG", "SIGHT GLASS"],
        "characteristics": "액위 측정용"
    },
    "TRANSMITTER": {
        "dat_file_patterns": ["PT_", "TT_", "FT_", "LT_"],
        "description_keywords": ["TRANSMITTER", "트랜스미터", "4-20MA"],
        "characteristics": "신호 전송용"
    },
    "INDICATOR": {
        "dat_file_patterns": ["PI_", "TI_", "FI_", "LI_"],
        "description_keywords": ["INDICATOR", "지시계", "DISPLAY"],
        "characteristics": "표시용"
    },
    "SPECIAL_INSTRUMENT": {
        "dat_file_patterns": ["INST_", "SPEC_"],
        "description_keywords": ["THERMOWELL", "ORIFICE PLATE", "MANOMETER", "ROTAMETER"],
        "characteristics": "특수 계기류"
    }
}

def classify_instrument(dat_file: str, description: str, main_nom: str) -> Dict:
    """
    간단한 INSTRUMENT 분류
    
    Args:
        dat_file: DAT_FILE 필드
        description: DESCRIPTION 필드
        main_nom: MAIN_NOM 필드 (연결 사이즈)
    
    Returns:
        간단한 계기 분류 결과
    """
    
    # 1. 재질 분류 (공통 모듈)
    material_result = classify_material(description)
    
    # 2. 계기 타입 분류
    instrument_type_result = classify_instrument_type(dat_file, description)
    
    # 3. 측정 범위 추출 (있다면)
    measurement_range = extract_measurement_range(description)
    
    # 4. 최종 결과
    return {
        "category": "INSTRUMENT",
        
        # 재질 정보
        "material": {
            "standard": material_result.get('standard', 'UNKNOWN'),
            "grade": material_result.get('grade', 'UNKNOWN'),
            "material_type": material_result.get('material_type', 'UNKNOWN'),
            "confidence": material_result.get('confidence', 0.0)
        },
        
        # 계기 정보
        "instrument_type": {
            "type": instrument_type_result.get('type', 'UNKNOWN'),
            "characteristics": instrument_type_result.get('characteristics', ''),
            "confidence": instrument_type_result.get('confidence', 0.0)
        },
        
        "measurement_info": {
            "range": measurement_range.get('range', ''),
            "unit": measurement_range.get('unit', ''),
            "signal_type": measurement_range.get('signal_type', '')
        },
        
        "size_info": {
            "connection_size": main_nom,
            "size_description": main_nom
        },
        
        "purchase_info": {
            "category": "완제품 구매",
            "supplier_type": "계기 전문업체",
            "lead_time": "2-4주",
            "note": "사양서 확인 후 주문"
        },
        
        # 간단한 신뢰도
        "overall_confidence": calculate_simple_confidence([
            material_result.get('confidence', 0),
            instrument_type_result.get('confidence', 0)
        ])
    }

def classify_instrument_type(dat_file: str, description: str) -> Dict:
    """계기 타입 분류"""
    
    dat_upper = dat_file.upper()
    desc_upper = description.upper()
    
    # DAT_FILE 패턴 확인
    for inst_type, type_data in INSTRUMENT_TYPES.items():
        for pattern in type_data["dat_file_patterns"]:
            if pattern in dat_upper:
                return {
                    "type": inst_type,
                    "characteristics": type_data["characteristics"],
                    "confidence": 0.9,
                    "evidence": [f"DAT_PATTERN: {pattern}"]
                }
    
    # DESCRIPTION 키워드 확인
    for inst_type, type_data in INSTRUMENT_TYPES.items():
        for keyword in type_data["description_keywords"]:
            if keyword in desc_upper:
                return {
                    "type": inst_type,
                    "characteristics": type_data["characteristics"],
                    "confidence": 0.8,
                    "evidence": [f"KEYWORD: {keyword}"]
                }
    
    return {
        "type": "UNKNOWN",
        "characteristics": "분류되지 않은 계기",
        "confidence": 0.0,
        "evidence": ["NO_INSTRUMENT_TYPE_FOUND"]
    }

def extract_measurement_range(description: str) -> Dict:
    """측정 범위 추출 (간단히)"""
    
    desc_upper = description.upper()
    
    # 압력 범위
    pressure_match = re.search(r'(\d+(?:\.\d+)?)\s*-\s*(\d+(?:\.\d+)?)\s*(PSI|BAR|KPA)', desc_upper)
    if pressure_match:
        return {
            "range": f"{pressure_match.group(1)}-{pressure_match.group(2)}",
            "unit": pressure_match.group(3),
            "signal_type": "PRESSURE"
        }
    
    # 온도 범위
    temp_match = re.search(r'(\d+(?:\.\d+)?)\s*-\s*(\d+(?:\.\d+)?)\s*(°?C|°?F)', desc_upper)
    if temp_match:
        return {
            "range": f"{temp_match.group(1)}-{temp_match.group(2)}",
            "unit": temp_match.group(3),
            "signal_type": "TEMPERATURE"
        }
    
    # 신호 타입
    if "4-20MA" in desc_upper or "4-20 MA" in desc_upper:
        return {
            "range": "4-20mA",
            "unit": "mA",
            "signal_type": "ANALOG"
        }
    
    return {
        "range": "",
        "unit": "",
        "signal_type": ""
    }

def calculate_simple_confidence(scores: List[float]) -> float:
    """간단한 신뢰도 계산"""
    valid_scores = [s for s in scores if s > 0]
    return round(sum(valid_scores) / len(valid_scores), 2) if valid_scores else 0.0